Table_1_Synergistic Effects of Climate and Land-Cover Change on Long-Term Bird Population Trends of the Western USA: A Test of Modeled Predictions.pdf

Climate and land-use change are predicted to lead to widespread changes in population dynamics, but quantitative predictions on the relative effects of these stressors have not yet been examined empirically. We analyzed historical abundance data of 110 terrestrial bird species sampled from 1983 to 2010 along 406 Breeding Bird Survey (BBS) across the western USA. Using boosted-regression trees, we modeled bird abundance at the beginning of this interval as a function of (1) climate variables, (2) Landsat-derived landcover data, (3) the additive and interactive effects of climate and land-cover variables. We evaluated the capacity of each model set to predict observed 27-year bird population trends. On average, 45 species significantly declined over the period observed and only 8 increased (mean trend = −0.84%/year). Climate change alone significantly predicted observed abundance trends for 44/108 species (mean 0.37 ± 0.09 [SD]), land-cover changes alone predicted trends for 47/108 species (mean r = 0.36 ± 0.09), and the synergistic effects predicted 59/108 species (mean r = 0.37 ± 0.11). However, for 37 of these species, including information on land-cover change increased prediction success over climate data alone. Across stressors, species with trends that were predicted accurately were more likely to be in decline across the western USA. For instance, species with high correlations between predicted and observed abundances (>r = 0.6) were declining at rates that were on average >2%/year. We provide the first empirical evidence that abundance models based on land cover and climate have the capacity to predict the species most likely to be at risk from climate and land-use change. However, for many species there were substantial discrepancies between modeled and observed trends. Nevertheless, our results highlight that climate change is already influencing bird populations of the western U.S. and that such effects often operate synergistically with land-cover change to affect population declines.

气候与土地利用变化被预测将引发种群动态的广泛改变，但目前尚无实证研究检验这两类胁迫因子的相对影响的定量预测结果。我们分析了1983年至2010年间，沿美国西部406条繁殖鸟类调查（Breeding Bird Survey, BBS）路线采样得到的110种陆生鸟类的历史种群丰度数据。借助提升回归树（boosted-regression trees），我们以如下变量为自变量，对该时段初期的鸟类种群丰度进行建模：(1) 气候变量；(2) 陆地卫星（Landsat）反演的土地覆被数据；(3) 气候与土地覆被变量的加性及交互效应。我们评估了每一组模型集对观测到的27年鸟类种群趋势的预测能力。平均而言，观测期内有45个物种的种群显著下降，仅8个物种呈增长趋势（平均趋势为-0.84%/年）。仅气候变化即可显著预测44/108个物种的观测丰度趋势（平均相关系数r=0.37±0.09，标准差SD）；仅土地覆被变化即可预测47/108个物种的种群趋势（平均r=0.36±0.09）；而二者的协同效应可预测59/108个物种的种群趋势（平均r=0.37±0.11）。然而，其中37个物种在纳入土地覆被变化的信息后，其预测精度相较于仅使用气候数据的模型得到了提升。综合各类胁迫因子来看，种群趋势被准确预测的物种，在美国西部的种群下降概率更高。例如，预测丰度与观测丰度间相关系数较高（r>0.6）的物种，其种群平均下降速率超过2%/年。我们首次提供实证证据表明，基于土地覆被与气候的丰度模型，能够识别出最易受气候与土地利用变化威胁的物种。不过，对于多数物种而言，模型预测结果与观测趋势之间仍存在显著偏差。尽管如此，我们的研究结果强调，气候变化已对美国西部的鸟类种群产生影响，且这类影响通常与土地覆被变化协同作用，共同推动种群下降。